如何并行执行python subTests？

Question

考虑下面的unittest.TestCase ，它实现了同一个测试的两个版本，唯一的区别是一个使用multiprocessing并行执行subTest 。

import multiprocessing as mp
from unittest import TestCase


class TestBehaviour(TestCase):
    def _test_equals(self, val):
        for target_val in [1, 2]:
            with self.subTest(target=target_val, source=val):
                self.assertEqual(val, target_val)

    def test_equality_parallel(self):
        with mp.Pool(processes=4) as pool:
            pool.map(self._test_equals, [1, 2])
            pool.join()
            pool.close()

    def test_equality(self):
        for val in [1, 2]:
            self._test_equals(val)

串行版本test_equality按预期工作并产生以下测试失败：

======================================================================
FAIL: test_equality (temp.TestBehaviour) (target=2, source=1)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "temp.py", line 11, in _test_equals
    self.assertEqual(val, target_val)
AssertionError: 1 != 2

======================================================================
FAIL: test_equality (temp.TestBehaviour) (target=1, source=2)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "temp.py", line 11, in _test_equals
    self.assertEqual(val, target_val)
AssertionError: 2 != 1

另一方面， test_equality_parallel会导致错误，因为TestCase不能被pickle：

Traceback (most recent call last):
  File "temp.py", line 15, in test_equality_parallel
    pool.map(self._test_equals, [1, 2])
  File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 364, in map
    return self._map_async(func, iterable, mapstar, chunksize).get()
  File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 768, in get
    raise self._value
  File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/pool.py", line 537, in _handle_tasks
    put(task)
  File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/connection.py", line 206, in send
    self._send_bytes(_ForkingPickler.dumps(obj))
  File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/multiprocessing/reduction.py", line 51, in dumps
    cls(buf, protocol).dump(obj)
TypeError: cannot pickle '_io.TextIOWrapper' object

现在我知道我可以将_test_equals拆分为类之外的独立函数； 但是，我想保留subTest的行为，以便更好地记录（和后续调试）测试失败。

如何并行运行测试，但保留此subTest功能？

更新

我也尝试过使用pathos.multiprocessing.ProcessingPool来规避TestCase序列化的问题； 然而，在这种情况下pool.join()引发ValueError: Pool is still running 。

from pathos.multiprocessing import ProcessingPool
...
    def test_equality_parallel(self):                                           
        pool = ProcessingPool(processes=4)                                      
        pool.map(self._test_equals, [1, 2])
        pool.join()

更新 2

这个问题肯定是相关的。 提出的第一个解决方案，为要从子进程调用的方法创建第二个类是不合适的，因为它不能使用subTest 。 第二个，从TestCase删除 unpickleable _Outcome对象似乎很subTests ，并且考虑到子进程正在运行subTests ，看起来也不合适。

Answer 1

我是pathos （以及dill和multiprocess ）作者。 您仍然看到跨进程的pathos序列化错误。 您可以尝试跨线程序列化。 线程并行可能适合这种级别的功能。

import multiprocess.dummy as mp
from unittest import TestCase


class TestBehaviour(TestCase):
    def _test_equals(self, val):
        for target_val in [1, 2]:
            with self.subTest(target=target_val, source=val):
                self.assertEqual(val, target_val)

    def test_equality_parallel(self):
        with mp.Pool(processes=4) as pool:
            pool.map(self._test_equals, [1, 2])
            pool.join()
            pool.close()

    def test_equality(self):
        for val in [1, 2]:
            self._test_equals(val)

其中产生：

======================================================================
FAIL: test_equality (test_equaltiy.TestBehaviour)
----------------------------------------------------------------------
...[snip]...
AssertionError: 1 != 2

======================================================================
FAIL: test_equality_parallel (test_equaltiy.TestBehaviour)
----------------------------------------------------------------------
...[snip]...
AssertionError: 1 != 2

----------------------------------------------------------------------
Ran 2 tests in 0.108s

FAILED (failures=2)

这告诉我您可以使用dill的序列化变体（即在dill.settings ）来解决序列化问题。 请参阅： https : //github.com/uqfoundation/multiprocess/issues/48 。